I seem to be having the same issue as someone in this thread:
http://www.keil.com/forum/15652/
I had it before and got it to go away, I'm not sure how...
Now nothing I do works, I cant debug at all...
I can still see the disassembly and step through that, but that isn't that useful when debugging data structures etc.
Does anyone have any clue what could cause this?
It happens when simulating or debugging on chip (STM32F4xx)
I tried deleting all output files and regenerating as well as everything listed in the above thread.
No errors are printed at all.
Any help would be appreciated
Did you turn off optimization?
There is typically not a one-to-one linear representation produced by the compiler, lines can be folded and done in orders that suit the compiler/cpu. You should look at the left column for different shades to see what lines can, and cannot be breakpointed. This is a good indicator if there is specific code behind the line, or not.
Without seeing the offending code it's hard to know exactly what you're working with or why it might pose difficulty.
I did try to turn off optimization, the problem isn't that code is being optimized away, I'm certain it's some sort of issue with the debugger. I can't step in c code at all but I see the disassembly in mixed mode, so it's not break points missing from optimizing. I can also step in the startup assembly file but when I jump to main it doesn't switch files at all, and if I open main there is no stepping arrow etc
you would not happen to have a few assembler lines in the c?
No, just the set up code and a little test program for my queue data structure
If someone were try to replication your issue what would they need?
You'd need to select a few peripherals in STM32CUBEMX, generate the code (simply generates startup code), set the processor to STM32F429II (though it happens no matter what processor is selected). Then try to debug with either the simulator or the debugger, my debugger is the ST-Link (I'm currently using the Disco board). It seemed to happen out of nowhere without changing a setting, very strange. I think other projects still work in the debugger so it is something isolated to this project somehow...thanks!
And are you using uVision 3 like the thread you cited, or something of more recent vintage? If so, what.
I've made progress!!! It seems that this only happens on my PC, which is a mac running windows in parallels...
I just installed uvision on my partner's regualar windows PC and I can step thorugh C code, but now it breaks on every statement even if there isn't a breakpiont there, any reason this could happen?
OK, I've narrowed it down...
When I compile on the mac (running parallels) it doesnt work in debug
When I compile on a standard PC, the the debug works on any computer.
This has to do with how the parallels mac outputs the debug information.
Any Ideas? It'd be a real pain to have to have my partner compile for me every time...
Have you compared all log output from a full build on your machine and the PC?
Same parameters? Same versions of binaries? No strange path used by your machine like possibly a name with space in it?